RAM with Fault model

i would appreciate someone's help

i have to design a VHDL model describing RAM with two different fault
models, such as stuck at fault and coupling fault
the following code is for genric RAM model, i do not know how can i
optimise it so that a functional fault is injected
library ieee;
use ieee.std_logic_1164.all;


entity RAM8X8 is
port ( Address : in integer range 0 to 7;
Data : inout std_logic_vector (7 downto 0);
WE : in std_logic;
OE : in std_logic);
CS : in std_logic);


Architecture behaviour of SRAM8X8 is

Type RAM8X8_array is array (0 to 7) of std_logic_vector (7 downto
0);

Begin
Process (address,Data,WE,OE,CS ) is
Variable mem: RAM8X8_array;

Begin
Data<= (others => ‘Z’);
if CS=’0’ then
if OE=’0’ then
Data <= mem(Address);
Elseif WE=’0’ then
Mem(address):= Data;
End if;
End if;
End process;
End RAM8X8;

Many Thanks
Allan

 

[KJ]

On Jun 24, 8:06 pm, o.tam...@hotmail.co.uk wrote:

i would appreciate someone's help

i have to design a VHDL  model describing RAM with two different fault
models, such as stuck at fault and coupling fault
the following code is for genric RAM model, i do not know how can i
optimise it so that a functional fault is injected

Adding Data(0) <= '0'; at the end of your process creates a stuck at
'0' model.

Add a new variable called 'Address_faulty' like the following will
create another stuck at '0' model.
address_faulty := address - (address mod 2);
replace Mem(address) with Mem(address_faulty);

KJ

 

[Marcus Harnisch]

Hi Allan

o.tamimi@hotmail.co.uk writes:

i have to design a VHDL model describing RAM with two different fault
models, such as stuck at fault and coupling fault

Since you care about static faults only, I would generate them as part
of the write action. Although generating the error on read has the nice
side effect that you can tell what should have been returned. OTOH,
this violates the rule that checkers should be separated from models...

BTW: There are three different coupling faults. Check which type you
are supposed to implement.

Regards
Marcus

--
note that "property" can also be used as syntaxtic sugar to reference
a property, breaking the clean design of verilog; [...]

(seen on http://www.veripool.com/verilog-mode_news.html)